11.1.4 Tensor Index Position Behavior Scope

Tensor Index Position Behavior Scope is the delineation of exactly what information the vertical placement of an index, whether raised as a superscript or lowered as a subscript, conveys about a tensor's transformation behavior, and the limits within which that placement alone can be relied upon to determine variance type.

What Index Position Determines

The Transformation Rule for That Slot

The primary function of index position is to specify, without any further annotation, which Jacobian factor a given slot uses under a change of basis: a superscript slot uses the direct factor and a subscript slot uses the inverse factor. This single visual cue is sufficient, within a properly defined tensor, to reconstruct the entire transformation law once all index positions in an expression are known.

Correct Pairing in the Summation Convention

Index position also determines which pairs of indices are eligible to be summed under the summation convention, since a repeated index must appear once as a superscript and once as a subscript; two occurrences of the same letter both as superscripts, or both as subscripts, do not constitute a valid summation pair.

Scope Within a Fixed Tensor

Position Is Meaningful Only Relative to a Declared Object

Index position conveys transformation behavior only in the context of a symbol already established to be a genuine tensor. Attaching a superscript to a quantity that has not been shown to obey the tensor transformation law does not, by itself, cause that quantity to behave contravariantly; the notation records an intended behavior, but the underlying object must independently satisfy the transformation rule for the notation to be accurate.

Consistency Across an Entire Expression

Within a single valid tensor equation, every occurrence of a given index letter belonging to the same tensor slot must retain the same vertical position throughout the derivation, since changing an index from a superscript to a subscript partway through a calculation, without an explicit index-lowering operation using the metric, silently changes which transformation rule is being claimed for that slot.

Boundary of What Position Alone Can Convey

Position Does Not Certify Tensorial Status

Because index position is a purely notational device, it cannot by itself guarantee that the object it decorates transforms homogeneously under a change of basis. Christoffel symbols, for example, are written with a mix of superscript and subscript indices resembling a mixed tensor, yet the object as a whole fails to obey the pure tensor transformation law, showing that index position describes an intended, not a guaranteed, transformation behavior.

Position Does Not Encode Which Specific Coordinate System Is in Use

Index position indicates only the variance type of a slot, not which particular coordinate system the components are expressed in; that additional information is carried separately, typically through a prime mark or a distinct index alphabet, layered on top of the vertical position rather than replacing it.

Position Alone Does Not Fix a Basis-Independent Meaning

An index position tells the reader how a component transforms, but the basis-independent geometric object being described requires the full data of the transformation law across the entire index array, together with the summation convention, taken as a whole; a single index's position considered in isolation does not by itself determine the tensor's invariant meaning.

Interaction With Index Raising and Lowering

Position Changes When the Metric Is Applied

When an index is deliberately raised or lowered using the metric tensor or its inverse, the change in vertical position reflects a genuine change in variance type, converting a covariant slot into a contravariant one or the reverse, and this operation is explicitly distinct from an unjustified notational shift, since it is accompanied by contraction with the metric.

Scope Limited to Spaces With a Metric

The ability to move an index between upper and lower position through raising and lowering is only available in a space equipped with a metric or an equivalent nondegenerate bilinear structure; on a bare vector space without such structure, the position of an index is fixed once the object is defined and cannot be reassigned, marking a further boundary on what index position behavior can achieve.